Searching \ for '[EE] Pull-up/down resistors' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=pull+updown+resistors
Search entire site for: 'Pull-up/down resistors'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Pull-up/down resistors'
2005\02\21@175231 by Julian Kain

picon face
Hi all,

I have a couple general and very simple questions regarding the use of
pull-up and pull-down resistors:

I'm currently in the schematic phase of a MP3 player system that uses
the Micronas MAS3507D chip -- it seems that every schematic I look at
has a different way of utilizing pull-up and pull-down resistors. As a
rule of thumb, when are they required? If I want to tie digital inputs
to "high" (VDD), do I _always_ need pull-up resistors to limit the
current? And, likewise, if I want to tie digital inputs to ground (VSS),
do I _always_ use pull-down resistors? Furthermore, what resistance
should I use? The schematics I'm referencing seem to use anywhere
between 2K and 100K, with several values in between. Is this an art?

Thanks so much for any insight you may be able to provide.

Julian

2005\02\21@183951 by Harold Hallikainen

face picon face
If an input is always going to sit high or low, I just tie it to Vcc or
Vss. If a switch is going to pull it in the opposite direction, I
generally use 10k.

Harold
(The ideal design has zero parts.)




--
FCC Rules Updated Daily at http://www.hallikainen.com

2005\02\21@201704 by Senior Design

picon face
>> If an input is always going to sit high or low, I just tie it to Vcc or Vss.

Thanks, but I'm confused as to the use of resistors often placed
between VDD (or VCC) and an input pin...

If a "pull-up" resistor is placed only between VDD and the line to be
switched, that makes sense. However, what then is a resistor simply
placed between VDD and the input? Is this a current-limiting resistor?
If so, what size is chosen?

You mention that one should simply tie the pin to VDD or VSS. But what
then of the many, many schematic examples that place a single resistor
on this line?

Thanks,
Julian

2005\02\21@202538 by csb

flavicon
face

> You mention that one should simply tie the pin to VDD or VSS. But what
> then of the many, many schematic examples that place a single resistor
> on this line?
'Tie' means 'via a resistor'. That means, if nothing else is acting on the
line, the 'default' value is either a 1 or a 0, depending on what the
resistor is tied to. for example, you might have a pull-down resistor
on an uC input. On the same line, you have a switch that closes the circuit
between the input and Vcc. Ohm's law. you then have a bit of current
flowing through the pull-down resistor to ground, and a very small bit
going through the input circuitry of the uC. It has nothing to do with
current limiting, except if it's placed serially along a signal, which
is used to reduce ringing due to fast rise/fall time.

Christian

2005\02\21@204702 by Rob Young

picon face


>>> If an input is always going to sit high or low, I just tie it to Vcc or
>>> Vss.
>
> Thanks, but I'm confused as to the use of resistors often placed
> between VDD (or VCC) and an input pin...
>
> If a "pull-up" resistor is placed only between VDD and the line to be
> switched, that makes sense. However, what then is a resistor simply
> placed between VDD and the input? Is this a current-limiting resistor?
> If so, what size is chosen?

Look at the specifications of the part, specifically what is the minimum
input current required to create a Vih on that pin?  Mr. Ohm is your friend.
But in general, for a +5V system, 10K to 100K is fine, just depends on how
"stiff" you want the pull-up (or down) to be.  If I'm feeling lazy or just
in a hurry, I keep a large supply of 47K resistors in various package types
around for prototypes and what-if's.  See more comments below.

>
> You mention that one should simply tie the pin to VDD or VSS. But what
> then of the many, many schematic examples that place a single resistor
> on this line?

Two reasons (and there are more which others will no doubt point out) but A)
if you decide later that line should have been at GND and you used a pull-up
resistor you can just ground the pin.  You burn a little power in the
resistor but makes for a quick modification in a prototype.  And B) if you
accidentally put +12V to your +5V (or +3V or whatever) rail your part might
survive (or it might blow the bond wires on its Vdd/Vss pins, a good
argument for current-limited supplies on the bench), at least long enough
for you to finish debugging and get replacement parts.  If you are fiddling
around and or just making a few of a particular design, the extra $0.03 to
pull-up lines vs tieing them to Vdd might be worth your time and money.

Size the resistor based on how much current will be drawn (and thus power
disipated) if a pulled-up line is brought low or a pulled-down line is
brought high.  Again, Mr. Ohm is your friend.  You will be balancing the
current required to produce a valid Vih or Vil vs overall power consumption
vs the driving devices ability to source or sink current vs power dissipated
in the pull-up/down resistor vs any inherent capacitance on a driving pin if
it changes from an output to open-collector/hi-z/input.  Lots of trade-offs.

Rob

2005\02\21@204942 by Jinx

face picon face

> Thanks, but I'm confused as to the use of resistors often placed
> between VDD (or VCC) and an input pin...

Check this out

http://www.piclist.com/techref/logic/xtrapins.htm


2005\02\21@211047 by Richard.Prosser

flavicon
face

If you can guarantee that the line will never be an output then tie it to
high or low and drive it appropriately.
Most of us have a bit of a feeling that "something" may go wrong enough to
make that pin temporarily
an output and so possibly draw or source damaging current.
On the other hand, if it is defined as an output and not connected to
anything then it may be
damaged by static or spikes etc. when powered off or during startup (before
it is configured)..
So unused pins are often connected to Vdd or Vss through a resistor. The
value of the resistor is dependent
on whether the pin is normally defined as an output or input (i.e a high
value or a low one (100k / 1k) and the level
of electrical noise expected to be received by the pin.

Richard P




>> If an input is always going to sit high or low, I just tie it to Vcc or
Vss.

Thanks, but I'm confused as to the use of resistors often placed
between VDD (or VCC) and an input pin...

If a "pull-up" resistor is placed only between VDD and the line to be
switched, that makes sense. However, what then is a resistor simply
placed between VDD and the input? Is this a current-limiting resistor?
If so, what size is chosen?

You mention that one should simply tie the pin to VDD or VSS. But what
then of the many, many schematic examples that place a single resistor
on this line?

Thanks,
Julian

2005\02\21@215426 by John J. McDonough

flavicon
face
----- Original Message -----
From: <spam_OUTRichard.ProsserTakeThisOuTspampowerware.com>
Subject: Re: [EE] Pull-up/down resistors


> value or a low one (100k / 1k) and the level
> of electrical noise expected to be received by the pin.

To some extent, this is the artsy part.

If I have an input pin, and maybe it's driving several other gates, I may
need a little lower value to drive all those gates.  But if it is only an
input, say pulled to ground by a pushbutton now and then, all I *really*
have to do is drive the PIC input pin.  Different pins have slightly
different specs, but they all require very little current.  From that
perspective, why not put a very high value there.  I could probably use a
couple of megs, or at least hundreds of K, and still get enough current to
raise the pin.

The problem is that the PIC input presents a very high impedance, and if the
rest of the line is also high impedance, then it will be easy for any nearby
lines to get that line to dance with it.  We generally don't like that, so
we go with a little more current.  Most of the time,  4.7K works just fine,
or 47K.  But if the line is short, and isn't near anything else, then higher
is fine, it saves a little power.  You can also work this from the other
side, which is why you see so many different choices.  If I can limit the
current in the nearby line, then it will have a harder time making this line
move.  So if I can keep everything in the vicinity to very low current, then
this line can work at very low current, too.

But as you can see, in 99% of the cases, the value is terribly non-critical.
Pretty much anything will work.  I happen to like 4.7K.  But if I'm already
in my 10K drawer, or my 47K drawer, I'll use that.  I gotta admit, if I'm
pulling the line to ground through a contact closure, I'll probably lean a
little higher, and if I'm doing this a bunch of times, I'll also lean toward
the high side, in the interest of power consumption.

Now, if I'm doing something to run off a battery, I might be a little more
careful.  A milliamp here and a milliamp there tends to add up.  But most of
the time, almost anything you lay your hands on will work just fine.

--McD


2005\02\21@221551 by Harold Hallikainen

face picon face


> You mention that one should simply tie the pin to VDD or VSS. But what
> then of the many, many schematic examples that place a single resistor
> on this line?
>

I don't believe the resistor is necessary. Perhaps someone who does feel
it is necessary can provide an explanation why. I've got several thousand
units in the field where an input is just tied high or low with no
problem. Inputs on chips normally have voltage limits of slightly above
Vcc and slightly below Vss. Vcc and Vss seem to be within these limits.

If, however, a line is bidirectional (could be an output, especially on
power up prior to initialization) a resistor is appropriate to limit
current during the time the line is being driven by the chip. If it's just
an input, though, I don't see a reason for the resistor. I'm open to
comment from others, though.

Harold



--
FCC Rules Updated Daily at http://www.hallikainen.com

2005\02\21@234204 by Rob Young

picon face
>
> I don't believe the resistor is necessary. Perhaps someone who does feel
> it is necessary can provide an explanation why. I've got several thousand
> units in the field where an input is just tied high or low with no
> problem. Inputs on chips normally have voltage limits of slightly above
> Vcc and slightly below Vss. Vcc and Vss seem to be within these limits.

In a production unit where you KNOW you will not need to be changing a
control input then tying directly to logic low or logic high is OK.  Build
enough units and those $0.03 resistors start to add up so if you can leave
them out, you save money on parts and assembly.

>
> If, however, a line is bidirectional (could be an output, especially on
> power up prior to initialization) a resistor is appropriate to limit
> current during the time the line is being driven by the chip. If it's just
> an input, though, I don't see a reason for the resistor. I'm open to
> comment from others, though.

On a prototype, especially if you haven't used a particular chip before, it
is nice to include the pull-up/pull-down resistor in case you change your
mind or discover you need to dynamically control a line you originally
though should have been staticly tied high or low.

Also the bit about bi-directional lines is appropriate if they are outputs
at power-on, to later be used as inputs.

Rob


>
> Harold
>
>
>
> --
> FCC Rules Updated Daily at http://www.hallikainen.com
> --

2005\02\22@004613 by Russell McMahon

face
flavicon
face
>> I don't believe the resistor is necessary. Perhaps someone who does
>> feel
>> it is necessary can provide an explanation why.

> In a production unit where you KNOW you will not need to be changing
> a control input then tying directly to logic low or logic high is
> OK.

>> If, however, a line is bidirectional (could be an output,
>> especially on
>> power up prior to initialization) a resistor is appropriate to
>> limit
>> current during the time the line is being driven by the chip. If
>> it's just
>> an input, though, I don't see a reason for the resistor. I'm open
>> to
>> comment from others, though.

In an ideal world what you set stays set. Inputs stay that way and
outputs do too. Circuits don't change with time, the code never needs
rewriting because it was perfect first time and the user spec was so
good it anticipated every need and no updates were needed . Ever.

In a real world things change. Murphy brings electrostatic discharge,
power supply glitches, power supply sags, battery spikes, .... .
IF a pin tied high or low changes to an output the processor will
sink/source appropriately. This may not matter at all. And it may
cripple the product. It MAY cause physical damage to the processor. It
MAY cause a battery to flatten far far sooner than it ought. And it
may not. Code or hardware changes may short the open or open the
short - the pin's unused so nobody may notice.

If an unused pin is set to output and left unconnected and Murphy
flips it to an input, all may be well. Current drain may not be
grossly affected and product operation may be just spiffing. But, the
product may fail, current drain may cause problems, processor MAY be
damaged (unlikely but possible), A2Ds may be noise up, oscillations
may occur and cause RFI. And more.

Whether such dangers are worth protecting against is a considered
engineering tradeoff. In many cases adding a high value resistor (1
megohm is usually OK) is a reasonable tradeoff.


       Russell McMahon



2005\02\22@075726 by olin_piclist

face picon face
csb wrote:
>> You mention that one should simply tie the pin to VDD or VSS. But what
>> then of the many, many schematic examples that place a single resistor
>> on this line?
> 'Tie' means 'via a resistor'.

Not unless it explicitly says so.  To "tie" one signal to another implies a
direct connection.  In some cases, it is prudent to put a resistor in
series, but that should be specified if so.

I see nothing wrong with directly connecting PIC I/O pins to Vdd or Vss.
Just make sure that the firmware doesn't screw up and set the pin as an
output and drive it the opposite way from the rail it's tied to.  However,
that is a bug that needs to be fixed anyway.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\02\22@093355 by alan smith

picon face
IF....you never have the requirement to be able to drive it the opposite way its tied....no issues on just tying to gnd or Vcc/Vdd.   IF you need to be able to switch the line.....and never want it floating....then use a pup or pdn
Alot comes from experiance

Harold Hallikainen <.....haroldKILLspamspam@spam@hallikainen.com> wrote:


> You mention that one should simply tie the pin to VDD or VSS. But what
> then of the many, many schematic examples that place a single resistor
> on this line?
>

I don't believe the resistor is necessary. Perhaps someone who does feel
it is necessary can provide an explanation why. I've got several thousand
units in the field where an input is just tied high or low with no
problem. Inputs on chips normally have voltage limits of slightly above
Vcc and slightly below Vss. Vcc and Vss seem to be within these limits.

If, however, a line is bidirectional (could be an output, especially on
power up prior to initialization) a resistor is appropriate to limit
current during the time the line is being driven by the chip. If it's just
an input, though, I don't see a reason for the resistor. I'm open to
comment from others, though.

Harold



-- FCC Rules Updated Daily at http://www.hallikainen.co

2005\02\22@121756 by Dwayne Reid

flavicon
face
At 10:46 PM 2/21/2005, Russell McMahon wrote:

>In a real world things change. Murphy brings electrostatic discharge,
>power supply glitches, power supply sags, battery spikes, .... .
>IF a pin tied high or low changes to an output the processor will
>sink/source appropriately. This may not matter at all. And it may cripple
>the product. It MAY cause physical damage to the processor. It MAY cause a
>battery to flatten far far sooner than it ought. And it may not. Code or
>hardware changes may short the open or open the short - the pin's unused
>so nobody may notice.

I see no problem with leaving unused I/O pins un-connected and set as output.

Some have mentioned the supposed problem of the pin floating for the few
microseconds at startup until the code gets around to setting the TRIS
registers.  I fail to see how that is a problem.

Others have mentioned the potential problem of an ESD event or other glitch
flipping a TRIS bit such that the un-used pin becomes an input.  While this
might become a problem if it were never dealt with, prudent design suggests
that all static registers (set only during initialization) should be
refreshed on a periodic basis.  After all, if a glitch might turn an unused
I/O pin into an input, who is to say that a similar glitch might not change
some other register that would actually affect operation?

In summary - easiest method of dealing with unused I/O pins is to leave
them unconnected and set as output.

dwayne

--
Dwayne Reid   <dwaynerspamKILLspamplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 20 years of Engineering Innovation (1984 - 2004)
 .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
    `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

2005\02\22@151550 by Russell McMahon

face
flavicon
face
> Others have mentioned the potential problem of an ESD event or other
> glitch flipping a TRIS bit such that the un-used pin becomes an
> input.  While this might become a problem if it were never dealt
> with, prudent design suggests that all static registers (set only
> during initialization) should be refreshed on a periodic basis.
> After all, if a glitch might turn an unused I/O pin into an input,
> who is to say that a similar glitch might not change some other
> register that would actually affect operation?

> In summary - easiest method of dealing with unused I/O pins is to
> leave them unconnected and set as output.

Reasonable comment - BUT you've made a good point and then lost it in
your summary. The summary that the beginner should have in their mind
after reading might better be: (capitalisation is only to make the
points - wouldn't be left in a  real summary).

   In summary - the EASIEST reasonably safe method of
   dealing with unused I/O pins is to leave  them
   unconnected and set as output AND all static registers
   should be refreshed on a periodic basis,

to which I'd add

   BUT the most prudent method is to set unused
   pins to inputs and connect them to ground or
   Vdd with a large value resistor.


       RM

2005\02\22@173036 by csb

flavicon
face
> > 'Tie' means 'via a resistor'.
> I see nothing wrong with directly connecting PIC I/O pins to Vdd or Vss.
In the case of an unused input, you're right. But the power supply won't
like it if you're using a switch with a zero ohm pull-up.

> Just make sure that the firmware doesn't screw up and set the pin as an
> output and drive it the opposite way from the rail it's tied to.  However,
> that is a bug that needs to be fixed anyway.
Agreed. Plus, for hobbyist/prototyping work, 1 resistor is NOT a big
cost just to make sure you haven't messed up your firm/hard/soft/tupper ware.

Christian

2005\02\23@035718 by ThePicMan

flavicon
face

>Others have mentioned the potential problem of an ESD event or other glitch flipping a TRIS bit such that the un-used pin becomes an input.  While this might become a problem if it were never dealt with, prudent design suggests that all static registers (set only during initialization) should be refreshed on a periodic basis.  After all, if a glitch might turn an unused I/O pin into an input, who is to say that a similar glitch might not change some other register that would actually affect operation?

I think that this risk is overstimated. If I thought otherwise, I'd set up a PIC with a special program and keep it powered until it reports such a "magical" change. But I think that under proper conditions (working power supply unit, no nuclear explosions nearby ;-) etc..) one could wait years without anything happening..

I fully agree that the proper way to deal with unused pins is to set them as output and leave them unconnected. If something else than the program changes the TRISx registers, then I think that we've got a much more serious problem anyway..

Greets,
TPM

2005\02\23@044939 by Jinx

face picon face
> I fully agree that the proper way to deal with unused pins is to set
> them as output and leave them unconnected. If something else than
> the program changes the TRISx registers, then I think that we've got
> a much more serious problem anyway..

The trouble is that the level of sophistication built into a product is
probably going to be optimal, perhaps for economic or practical
reasons

If someone happens to place one of your control boxes right next to
a radio transmitter or a noisy relay bank, which you might not have
foreseen them doing, and it goes ape, do you then have to think about
putting "Do not put next to transmitters or relay banks" stickers on ?
Or blow the budget by changing from a plastic to a metal box ? Or
add things like tying resistors so unwanted external noise can't make
it go ape ?

Obviously every product is different in its design, cost restrictions and
how important it is for it not to go wrong and the consequences if it
does, but it pays to at least consider risk management. Both of the above
instances happened to my products. Recalling units and adding literally
a couple of cents worth of resistors and a bit of aluminium foil made
them so much less susceptible to future problems and, touch wood,
they've all behaved themselves for some time now

To quote a list member's signature, which I keep in the back of my mind
with everything I do,

"There's never enough time to do it right - there's always enough time
to do it again"

2005\02\23@051024 by ThePicMan

flavicon
face
At 22.49 2005.02.23 +1300, you wrote:
{Quote hidden}

Ok, but specifically, why should this RFI only change TRISx and not a
GPR with even more catastrophic consequences?

That's why I say that unused pins can be left unconnected and set as
outputs.. because if TRISx changes, chances are that a lot of even more
"delicate" GPR's have changed already.

Metaphorically, what's the real usefulness of a parachute in a submarine
whose only serious risk is to sink? Sure, the parachute will help in case
an UFO picks up the submarine and then drops it from 10000 ft altitude..
but I think that adding a resistor to every unused pin is almost as
useless as above..

On a 18F8720 we have almost 4K of GPR's that may be changed by EMI/RFI,
and only a few of TRISx registers. I would concentrate my efforts for
reliability elsewhere than on the lone TRISx..

And about the firmware/software, well.. while I don't think to produce
bugless code all the time, I'm extremely confident that I won't write
a routine that changes TRISx improperly for no reason..

One thing I've always believe in, regardless of TRISx issues, is to
debug code also *before* you notice any bug.

2005\02\23@054020 by Jinx

face picon face
> Ok, but specifically, why should this RFI only change TRISx and
> not a GPR with even more catastrophic consequences?
>
> That's why I say that unused pins can be left unconnected and set
> as outputs.. because if TRISx changes, chances are that a lot of
> even more "delicate" GPR's have changed already.

I'm not trying to dodge that issue, but it really depends on the actual
application. Yes, it's quite possible that a GPR affected might be
something like a shadow port register which could have serious
consequences. I agree that some applications should be hardened
against outside interference. For something like a clock it hardly
matters - you just don't put the clock in a particular place or it goes
funny or chews through batteries if an o/p changes to a floating i/p

2005\02\23@065740 by Martin Tedjawardhana

picon face
This is an interesting read to "refresh" everyone's memory

http://www.seattlerobotics.org/encoder/mar97/basics.html


On Mon, 21 Feb 2005 17:52:31 -0500, Julian Kain <.....seniordesignKILLspamspam.....gmail.com> wrote:
{Quote hidden}

> -

2005\02\23@071532 by Russell McMahon

face
flavicon
face
>> Ok, but specifically, why should this RFI only change TRISx and
>> not a GPR with even more catastrophic consequences?
>>
>> That's why I say that unused pins can be left unconnected and set
>> as outputs.. because if TRISx changes, chances are that a lot of
>> even more "delicate" GPR's have changed already.

It's a matter of precautionary engineering. Nothing catches
everything.

In the case of pins which are used, if it's meant to be an output and
it has almost any sort of load, then changing to an input with not
directly cause damage, excess current drain, noise or worse. These may
occur as a second order effect due to the load now not being driven -
this very much depends on the load. Also, if the load is high /
minimal (say 1Mohm up) then the input may cause problems.

If an input changes to an output it may cause problems if the drive is
low impedance BUT if one wanted to one could almost always engineer
against that possibility. eg an A2D input gains from being driven from
a low impedance source, but an impedance high enough to not do any
damage of the pin becomes an output will also usually not affect A2D
performance adversely.

Also, pins that are utilised are far less likely to be overlooked when
changes are made at some time after initial design, whereas unused
pins have no "champion" and if something happens to change their
status it may not be noticed. For example, a port may contain a
certain mix of inputs and outputs and have a DDR mask set accordingly.
Some pins may be unused and set as outputs and left unconnected or as
inputs and connected to say Vdd. If at some later stage the DDR mask
is changed to reflect some new usage of the existing pins the change
may also affect the unused pins. eg "OK, all pins that I'm using on
that port are now outputs so I'll write $FF to the DDR". Wave goodbye
to the low current drain you were seeing!

All the arguments I have heard against treating unused pins careful,
both in this discussion and previously, have been of the form "you
don't need to take this level of precaution because everything else
that can happen is worse / we know what we are doing/ it's nit
important to us / ...". The only argument I've heard that addresses an
ADVANTAGE of not hardware treating the pins is that of cost and/or
space. ALL these arguments have a varying degree of validity depending
on circumstance but it MUST be recognised that what is being done is
trading off engineering technical excellence with other factors. Doing
so may be entirely valid as long as one realises the implications.

       RM



2005\02\23@083528 by olin_piclist

face picon face
Russell McMahon wrote:
> Also, pins that are utilised are far less likely to be overlooked when
> changes are made at some time after initial design, whereas unused
> pins have no "champion" and if something happens to change their
> status it may not be noticed. For example, a port may contain a
> certain mix of inputs and outputs and have a DDR mask set accordingly.
> Some pins may be unused and set as outputs and left unconnected or as
> inputs and connected to say Vdd. If at some later stage the DDR mask
> is changed to reflect some new usage of the existing pins the change
> may also affect the unused pins.

This issue can be dealt with by good software design.  I use /INBIT and
/OUTBIT preprocessor directives to declare all I/O pins.  Any pins not
explicitly declared are automatically set as outputs driven low by the
standard PORT module.  If pins are moved around or added or deleted, the
/INBIT and /OUTBIT pin declarations have to change by necessity.  I would
have to go out of my way to do something stupid for unused pins not to be
handled correctly automatically.

I've done somewhere between 50 and 100 PIC projects and never had a problem
with this.  A few times it's been handy to connect something to a previously
unused pin after the fact.  On designs that aren't that cost sensitive, I
usually bring each unused I/O pin to its own labeled pad to aid rework in
case it becomes necessary.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\02\23@184753 by Julian Kain

picon face
Hi all,

I was the one who originally started this thread... thank you each so
much for all of your feedback. I've learned a lot, to say the least.

With the knowledge gained, I believe I've finished up my schematic. This
is my very first schematic-through-PCB design, and there's a lot riding
on its success (like, for example, my graduation). I don't know if it's
"proper" to solicit reviewing help on this list, and please accept my
apologies if it's not. With that said, if anyone would be willing to
_quickly_ glance over my schematic and point out any _obvious_ errors or
suggestions, I would be so grateful. It's located at
http://www.juliankain.com/schematic_20050223.pdf

Thanks a lot for any specific design or otherwise general advice any of
you may have to lend.

Sincerely,
Julian Kain

2005\02\23@191403 by Rob Young

picon face
> http://www.juliankain.com/schematic_20050223.pdf
>
> Thanks a lot for any specific design or otherwise general advice any of
> you may have to lend.

1) LM340 5V regulator would probably like a bit more "bulk" capacitance
right at its output pin.  Add about 10uF.  The voltage rating should be at
least 10V.

2) Add an anti-reversal diode (1N4001 or similar) where your power enters
the board.  Yes you will need to compensate at least another 1V more input
voltage to keep the regulator headroom but you won't be as likely to
accidentally hook up power backwards...

3) Consider diode (ANODE at Vout, CATHODE at Vin) on the regulators to
guarantee their outputs don't swing far from the input voltages.  This is
more important if you end up adding a lot of output capacitance to the
regulators.

4) The 20MHz xtal you have picked, am I to assume the ground connections you
show are to ground the xtal's case?  If it isn't an xtal but really a 20MHz
oscillator then you don't need C12 and C13 and you would only be connecting
to the OSC-IN pin on the PIC.

5) I didn't spend a lot of time scrolling around to find out if you have
seperate analog and digital grounds and if you are connecting them at a
single point.  This helps to keep digital ground currents from making your
analog grounds flinch and flail.

6) For your MAX232 you show 1uF caps.  If you are using polarized caps (but
you show non-polarized caps so maybe this comment isn't necessary) observe
the polarity of the C+ and C- terminals etc when connecting the capacitors.

Like I said in #5 I didn't spend a lot of time scrolling around to check
connections.  My personal preference is for heirarchical schematics on 8.5 x
11 sheets with the top sheet being a block diagram and each following sheet
a small portion of the schematic, logically partitioned so that each sheet
stands nearly on its own for viewing and debugging.  Just personal
preference though...

Rob Young

2005\02\23@195344 by olin_piclist

face picon face
Julian Kain wrote:
> With the knowledge gained, I believe I've finished up my schematic. This
> is my very first schematic-through-PCB design, and there's a lot riding
> on its success (like, for example, my graduation). I don't know if it's
> "proper" to solicit reviewing help on this list, and please accept my
> apologies if it's not. With that said, if anyone would be willing to
> _quickly_ glance over my schematic and point out any _obvious_ errors or
> suggestions, I would be so grateful. It's located at
> http://www.juliankain.com/schematic_20050223.pdf
>
> Thanks a lot for any specific design or otherwise general advice any of
> you may have to lend.

There was way too much crammed onto one sheet.  I couldn't read it, even
though I have a 19" monitor set to 1600 x 1200 pixels.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\02\23@203237 by John J. McDonough

flavicon
face
----- Original Message -----
From: "Olin Lathrop" <EraseMEolin_piclistspam_OUTspamTakeThisOuTembedinc.com>
Subject: Re: [EE] Pull-up/down resistors


> There was way too much crammed onto one sheet.  I couldn't read it, even
> though I have a 19" monitor set to 1600 x 1200 pixels.

At least it was organized into nice logical chunks.

But why on earth would anyone specify 1%ers for pull-ups?

--McD


2005\02\23@204713 by Robert Rolf

picon face

Olin Lathrop wrote:

> Julian Kain wrote:
>
>> With the knowledge gained, I believe I've finished up my schematic. This
>> is my very first schematic-through-PCB design, and there's a lot riding
>> on its success (like, for example, my graduation). I don't know if it's
>> "proper" to solicit reviewing help on this list, and please accept my
>> apologies if it's not. With that said, if anyone would be willing to
>> _quickly_ glance over my schematic and point out any _obvious_ errors or
>> suggestions, I would be so grateful. It's located at
>> www.juliankain.com/schematic_20050223.pdf
>>
>> Thanks a lot for any specific design or otherwise general advice any of
>> you may have to lend.
>
>
> There was way too much crammed onto one sheet.  I couldn't read it, even
> though I have a 19" monitor set to 1600 x 1200 pixels.

It's a PDF. It SCALES beautifully in Adobe. 400x is quite readable,
but you have to pan and scan to see it all.

I just picked up a laserjet 4V (does 11 x 17") so it will be fun
to test the printer with a suitable document.

R


2005\02\23@205025 by Julian Kain

picon face
> There was way too much crammed onto one sheet.  I couldn't read it,
> even though I have a 19" monitor set to 1600 x 1200 pixels.

Yeah, I know -- it's pretty cramped. Sorry about that, but my student
version of Multisim doesn't allow for multi-page designs (actually, it
doesn't allow for off-page connectors). I did my best to break the
design up into logical blocks with a PIC interface for each.

In addition, the document is a vectorized PDF, so you can zoom in
without losing quality.

Thanks everyone,
Julian

2005\02\23@205307 by Julian Kain

picon face
> But why on earth would anyone specify 1%ers for pull-ups?

Could you please elaborate on this point? I picked these simply because
they were the available 10k resistors within the software -- I have no
particular commitment to them. I just need 1k's and an 0805 or 0603
footprint to design the PCB from.

Thanks,
Julian

2005\02\23@210157 by Julian Kain

picon face
Thanks very much, Rob. I appreciate the comments.

To respond to your #1 suggestion... the LM340 data sheet specifies the
capacitors I chose (.22uF and .1uF) as required in the shown "typical
application" circuit for fixed output regulator. In fact, it says that
the output capacitance is not required at all, and recommends a 0.1uF to
"help transient response... if needed." Could you please explain your
recommendation so that I better understand its application?

Thanks again,
Julian


Rob Young wrote:

{Quote hidden}

2005\02\23@212124 by Marcel Duchamp

picon face
Julian Kain wrote:
>  > But why on earth would anyone specify 1%ers for pull-ups?
>
> Could you please elaborate on this point? I picked these simply because
> they were the available 10k resistors within the software -- I have no
> particular commitment to them. I just need 1k's and an 0805 or 0603
> footprint to design the PCB from.
>
> Thanks,
> Julian

Good question.

For Julian, the question is about the resistor tolerance of parts used
for pullups.

These pullups could probably work with values from less than 1K to
greater than 100K.  So tolerance isn't much of an issue.

The likely reason someone would ask is than in the bad old days, 1%
resistors cost much more than 5%.  Heck, we even had 10% and 20%
resistors available.  But... this ain't the bad old days any more.  I
just checked the place I buy passives from and in 1000 qty, the price is
the same for 1% and 5%: about US$0.014 each.  So it seems it doesn't
make much difference.

I use 1% exclusively.  Here's why.  I send a kit of parts to an assembly
house to be loaded.  It's 100% surface mount.  If I have a critical part
 that MUST be 1%, then I will have extra work verifying they loaded 1%
parts there and didn't mix them up with 5% parts.  If 1% parts are used
everywhere, there is no chance of a mix up.  Easier for the assembly
folks, easier for me.

YMMV
MD

2005\02\23@212545 by Jose Da Silva

flavicon
face
On Wednesday 23 February 2005 04:53 pm, Olin Lathrop wrote:
> Julian Kain wrote:
> > With the knowledge gained, I believe I've finished up my schematic. This
> > is my very first schematic-through-PCB design, and there's a lot riding
> > on its success (like, for example, my graduation). I don't know if it's
> > "proper" to solicit reviewing help on this list, and please accept my
> > apologies if it's not. With that said, if anyone would be willing to
> > _quickly_ glance over my schematic and point out any _obvious_ errors or
> > suggestions, I would be so grateful. It's located at
> > www.juliankain.com/schematic_20050223.pdf
> >
> > Thanks a lot for any specific design or otherwise general advice any of
> > you may have to lend.
>
> There was way too much crammed onto one sheet.  I couldn't read it, even
> though I have a 19" monitor set to 1600 x 1200 pixels.

I like to see a fair bit all at once, but I have to agree that quite a few
lines are quite a-"maze"-ing <yuck, yuck, yuck>  ;-)

You may want to put the chips closer together so we don't follow the maze of
wires so much.

You appear to have made use of the gnd symbol
 |
----
\  /
\/

See if you could several power lines to similar symbol connectors such as:
+12v
----
 |

or other similar common voltages like +5v, etc.

another possible reduction if you can know how to work it is the "bus" line.
If a technician is to follow the path from your compact flash d0x to your PIC
pins PTxx they could easily lose their way and end up pointing at the wrong
pin when they point at the other chip.

The "bus" looks similar to:
-\   /-
-\   /-
-\___/-

bus lines should clean up your page fairly fast.
U1B.RH6,RH2,RF7/RF5 could be 2 buses, or you can group them as 2 wires closer
together separated by a bigger space, then 2 wires close together, example
see grid point Q1...Q10 so that the reader sees something in-common versus
what appears as 5 wires running in parallel.
If you can put those connections on the right-side of U1B and the left side
of U6, it further reduces the maze of wires.

The "No-connections" is best placed on the side or near the edges instead of
in the center of the page. We usualy try to group items and put important
stuff in the middle.

Your page identification is going to look very small plus you have a lot of
blank area towards the right.
"Radio Frequency...."
If you make use of the blank wasted space you could enlarge a fair bit of the
text within the box.
This box is best placed along an edge such as the bottom edge or right edge.

Are you actually running 2 separate gnds on your circuit board?
Just asking, because you appear to have made use of both symbols, and looking
at the way they are being used, I don't think you seem to understand the
reasons for separating gnds.  In any case, some of the gnds are being used
incorrectly considering how you have wired some of the VCCs, so check with
your teacher.

Few people like reading data that crosses over or under drawings, for example
see R25, R27, R30.....etc. Turn the writing 90degrees if it doesn't fit.

Unless you have a super-duper-regulated power supply, I would probably add a
10uf or better behind J2, otherwise you're going to have problems.

U1D.INT1 and U1D.INT2 could probably be used for something else, but you need
to verify that, not me.

Hopefully that should get you started.

2005\02\23@212853 by Jinx

face picon face

> > But why on earth would anyone specify 1%ers for pull-ups?
>
> Could you please elaborate on this point? I picked these simply
> because they were the available 10k resistors within the software --
> I have no particular commitment to them. I just need 1k's and an
> 0805 or 0603 footprint to design the PCB from

Unless the MAS3507D specifies 10.00k (had a quick look in its
d/s and didn't see anything) a general purpose 10k 5% would do.
You could probably rationalise and use 15k as elsewhere in the
circuit and 470R for the pulldown. But from what I gather at a glance,
most of those pins have dual functions as either inputs or outputs
and it seems unusual to have both a pull-up and a pull-down. IOW
in your circuit the pin is being set at 1/11th of the supply. I'm guessing
that each pin really needs just either a p/u or p/d to set the functionality

Or have you included the resistors to reserve a space for them on the
PCB and half won't actually be installed ?


2005\02\23@213520 by Jose Da Silva

flavicon
face
On Wednesday 23 February 2005 05:47 pm, Robert Rolf wrote:
{Quote hidden}

most people testing or repairing or working on your board will be using a
paper copy of your schematic. A computer and monitor is a lot of bulk on a
workbench and takes up too much valuable space from a manufacturing
technician's point of work.
Try reading your schematic through the bottom of a drinking glass, because
after you place a few too-many pc boards on those brand-new pages or get it
folded once or twice, your nice crisp lines are going to look rather
tattered  ;-/

2005\02\23@215155 by Julian Kain

picon face
> and 470R for the pulldown.

I did some quick calculations and came up with "less than 2.2k" for the
pull-downs, so I used 1k. Is there any particular reason why you mention
470?


> most of those pins have dual functions as either inputs or outputs
and it seems unusual to have both a pull-up and a pull-down.

Well, the really don't. Some have pull-ups and some have pull-downs. A
lot of those pins on the MAS3507D are for configuration, and I don't
want to completely commit to tieing them to high or low (where this
thread originated from, in fact). So I either tie them to high or low
depending on how I _expect_ to configure them, with the option to set
them differently via the PIC as necessary. Per my schematic, I believe
this is correct -- am I right in saying so?


> IOW in your circuit the pin is being set at 1/11th of the supply.

I don't quite understand this... could you please explain?


> I'm guessing that each pin really needs just either a p/u or p/d to
set the functionality
> Or have you included the resistors to reserve a space for them on the
PCB and half
> won't actually be installed ?

Functionality... with the PIC control as a backup in case functionality
needs to be changed.

Thanks!
Julian

2005\02\23@215713 by Julian Kain

picon face
Thanks for your suggestions, Jose. A few questions...

> Are you actually running 2 separate gnds on your circuit board?

Yes, analog and digital.


> [...] I don't think you seem to understand the reasons for separating gnds.

Well, I _thought_ I did. :-)  What am I doing wrong? I was making the assumption that digital components grounds (VSS) connect with the solid DGND symbol, and analog grounds (AVSS, ground of MAX232's ground, etc) connect to the standard AGND symbol. Did I make a mistake or do I have a fundamental misunderstanding?

> U1D.INT1 and U1D.INT2 could probably be used for something else

These interrupts need to detect received data from their respective RX lines (RX1 to INT1, RX2 to INT2), so they should be tied directly in with those RX lines.


Thanks!
Julian

2005\02\23@220352 by Jinx

face picon face
> I did some quick calculations and came up with "less than 2.2k" for
> the pull-downs, so I used 1k. Is there any particular reason why you
> mention 470?

Only that you're using them with the MAX232.  If you can do the
whole thing with 10k / 1k) or 15k / 470R that does make purchases
a lot simpler, possibly cheaper if you can conglomerate into a 10+,
25+, whatever break point, and of course lessens the chance of a
stuffing mistake. None of the resistor values appear to be critical,
more ballpark

2005\02\23@221418 by John J. McDonough

flavicon
face
----- Original Message -----
From: "Marcel Duchamp" <marcel.duchampspamspam_OUTsbcglobal.net>
Subject: Re: [EE] Pull-up/down resistors


> The likely reason someone would ask is than in the bad old days, 1%
> resistors cost much more than 5%.  Heck, we even had 10% and 20%
> resistors available.  But... this ain't the bad old days any more.  I
> just checked the place I buy passives from and in 1000 qty, the price is
> the same for 1% and 5%: about US$0.014 each.  So it seems it doesn't
> make much difference.

I have to admit, I was surprised.  It's been a long time since we had 10 and
20% resistors, but for quite a while 5%ers were dirt cheap, and 1% was like
gold.  I looked at my source and while I didn't see anyplace where the price
was the same, there was never much of a difference, and in some sizes, 1%
was cheaper.  Kind of a surprise.  But I did notice, too, that 5% have
gotten quite a bit more expensive since the last time I paid attention.

--McD


2005\02\23@222106 by John J. McDonough

flavicon
face
----- Original Message -----
From: "Jinx" <@spam@joecolquittKILLspamspamclear.net.nz>
Subject: Re: [EE] Pull-up/down resistors


> a lot simpler, possibly cheaper if you can conglomerate into a 10+,
> 25+, whatever break point, and of course lessens the chance of a
> stuffing mistake. None of the resistor values appear to be critical,
> more ballpark

In fact, do the math for 100.  Sometimes 100 resistors cost less than 5 or
6.  (I sort of guessed you're not making 10,000 copies of this circuit).

72/73 de WB8RCR    http://www.qsl.net/wb8rcr
didileydadidah     QRP-L #1446 Code Warriors #35


2005\02\23@225303 by Russell McMahon

face
flavicon
face
> To respond to your #1 suggestion... the LM340 data sheet specifies
> the capacitors I chose (.22uF and .1uF) as required in the shown
> "typical application" circuit for fixed output regulator. In fact,
> it says that the output capacitance is not required at all, and
> recommends a 0.1uF to "help transient response... if needed." Could
> you please explain your recommendation so that I better understand
> its application?

An ideal regulator needs no output capacitance at all as it responds
to all load changes up to its rated limit.
A less than ideal regulator requires "some" to provide current while
the regulator 'gets its act together' on very rapid changes. With very
old regulators it may have been useful to add a substantial reservoir
capacitor although doing so can have some disadvantages as well. The
LM340 is new enough (although very old) to only need such for fairly
spiky loads.

Some modern regulators need larger output capacitors for stability
purposes - without them they will oscillate under certain load
conditions. The LM340 is not in this category. In some (usually
modern) regulators the output capacitor must have an ESR (equivalent
series resistance) that is neither too high OR too low - too "good" an
output cap can also cause oscillations! The LM340 is also not in this
category.

If the regulator input voltage drops suddenly the output capacitor(s)
will attempt to discharge back into the regulator. In some cases the
energy in the capacitor can damage the regulator if precautions are
not taken to prevent this discharge. The LM340 IS in this category -
to protect it against large output capacitors a usually reverse biased
diode should be connected from output to input. Capacitors of a few
microfarad are usually OK BUT see the data sheet. Input voltage can
fall rapidly if heavily loaded by other sources when the power is
removed OR due to a fault short circuit on the input side. Worst case
a regulator may fail such that it subsequently allows input voltage
through when power is next applied so its worth not letting them fail
this way - the internal thermal etc protection does not protect
against such failures. Some regulators are protected against such
failures. The LM340 is not in this category :-).

The LM293x family is probably a better choice of regulator than the
LM340. It has protection against most problems and better drop out
performance (nominally its an LDO regulator). Some members of the
LM293x family are drop in replacements for an LM340. Given the cost of
the rest of the circuitry, a very small amount extra spent on a few
LM2931's or similar will be well spent. Note that they have their own
output capacitor requirements - see data sheet.

An oscillating regulator can be a great mystery. The mean output
voltage is about correct and a meter will tell you that all is OK. A
peek with an oscilloscope will horrify you! ALWAYS check things with a
scope when it matters. Similarly, regulator dropout due to AC ripple
on the input and marginal operation may not be obvious but can wreak
havoc - especially with analog circuits. An LDO makes it less likely
that this will happen in any given case. (I have made an LM340
oscillate but it usually takes special skill :-) ).





       RM


2005\02\23@225524 by Jinx

face picon face
> In fact, do the math for 100.  Sometimes 100 resistors cost
> less than 5 or 6

Sometimes that bulk pricing can cause you decision-making problems
too. I recently needed 300 SMT resistors. Now, the reel price is about
0.128c each which is very tempting. But I'd have 4700 left over, basically
dead money, most suppliers dealing with only whole reels. Eventually I
got them from a place that splits reels for a business and paid 1.3c. Still
a good price considering the quantity and sure better than the 11c out of
a mail order catalogue. There was a moment when I very nearly said yes
to the whole reel, purely on price per resistor, but common sense prevailed

2005\02\23@225935 by Russell McMahon

face
flavicon
face
>> Are you actually running 2 separate gnds on your circuit board?
>
> Yes, analog and digital.

> Well, I _thought_ I did. :-)  What am I doing wrong? I was making
> the assumption that digital components grounds (VSS) connect with
> the solid DGND symbol, and analog grounds (AVSS, ground of MAX232's
> ground, etc) connect to the standard AGND symbol. Did I make a
> mistake or do I have a fundamental misunderstanding?

As long as they connect to each other somewhere - preferably at a
single point. If you can manage an infinite extent, zero impedance
ground plane then you can connect to it anywhere. As you usually
can't,  then star grounding and ground separation has its place.
people will now howl with horror and a holy war will ensue  :-)



       RM


2005\02\23@225936 by Russell McMahon

face
flavicon
face
> 72/73 de WB8RCR    http://www.qsl.net/wb8rcr
> didileydadidah     QRP-L #1446 Code Warriors #35

Non sequitur

Da di di dit
dit
dit
di di dah dit
...
Dit
di di dit
di di dit
dit
dah dit
dah di dah dit
dit
.

Beef Essence.
:-)

Sounds lovely when sent :-)

       RM


2005\02\23@231812 by Gaston Gagnon

face
flavicon
face
Julian Kain wrote:
>  > There was way too much crammed onto one sheet.  I couldn't read it,
>  > even though I have a 19" monitor set to 1600 x 1200 pixels.
>
> Yeah, I know -- it's pretty cramped. Sorry about that, but my student
> version of Multisim doesn't allow for multi-page designs (actually, it
> doesn't allow for off-page connectors). I did my best to break the
> design up into logical blocks with a PIC interface for each.

Hi Julian,
Here suggestions to remove clutter and help make the schematic more
readable:

1) First remove all chip supplies. What is left is usually required to
understand the logic of the system;

2) Grouped together, possibly on a different sheet, all regulators,
chips supplies and bypass capacitors;

3) Draw similar signals as busses.

I thing an excellent example of this technique is demonstrated by the
hexapod.sch distributed as example with Eagle cad software.
I have created a pdf of this file and sent it off list (as it is a
voluminous 400k file).

Hope this help.
Gaston

2005\02\23@232044 by Dwayne Reid

flavicon
face
At 12:50 AM 2/23/2005, ThePicMan wrote:

> >Others have mentioned the potential problem of an ESD event or other
> glitch flipping a TRIS bit such that the un-used pin becomes an
> input.  While this might become a problem if it were never dealt with,
> prudent design suggests that all static registers (set only during
> initialization) should be refreshed on a periodic basis.  After all, if a
> glitch might turn an unused I/O pin into an input, who is to say that a
> similar glitch might not change some other register that would actually
> affect operation?
>
>I think that this risk is overstimated. If I thought otherwise, I'd set up
>a PIC with a special program and keep it powered until it reports such a
>"magical" change. But I think that under proper conditions (working power
>supply unit, no nuclear explosions nearby ;-) etc..) one could wait years
>without anything happening..

Yeah - I used to think that way as well.  That changed when one of my very
early prototypes of a heating controller (based on 16c71) faulted in an
unsafe condition - presumably from an external glitch.  I found that I
could fault the unit at will by inducing external transients.

All other development stopped until I found out how to eliminate those
unsafe conditions.  I learned a lot over the succeeding several weeks -
such that I was NOT able to glitch the device into a unsafe state.  Not
even with a piezo sparker jumping a quarter inch into various nodes.

Part of the strategy turned out to be constant refresh of static
registers.  Customers like it a lot better if your device can recover
automatically from a transient fault if possible.  In fact, that simple
step eliminated the failures that I first noticed.

But that wasn't good enough for this particular project.  I wound up adding
extra hardware to verify safe conditions: hardware comparitors to check
analog levels, an external watchdog timer that disables relay drivers as
well as forcing a processor reset, several other things.  Basically, it was
set up so that even if the processor hung in a bad state, the external
hardware could disable the controller.

I've killed a lot of units with ESD testing but never had a unit fail unsafe.

And now I always refresh registers - even on the simplest 12c508 projects.

dwayne

--
Dwayne Reid   <KILLspamdwaynerKILLspamspamplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 21 years of Engineering Innovation (1984 - 2005)
 .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
    `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

2005\02\23@233204 by Robert Young

picon face
----- Original Message -----
From: "Julian Kain" <RemoveMEseniordesignTakeThisOuTspamgmail.com>
To: "Microcontroller discussion list - Public." <spamBeGonepiclistspamBeGonespammit.edu>
Sent: Wednesday, February 23, 2005 8:01 PM
Subject: Re: [EE] Pull-up/down resistors


{Quote hidden}

Russell made some good points about linear regulators in general.  In
summary, just don't blindly follow the sample circuit on the first page of
the datasheet.  Linear regulators seem like trivial things with their 3
(sometimes 4, 5, 6 or more) pins and most of the time they work great.  But
they only work so well because LOTS of thought went into their designs.

My suggestion for adding some bulk capacitance was a knee-jerk reaction to
seeing a linear regulator (any family) on a schematic and not seeing a
reasonably large cap near by.  Since this is going to be a one-off design, I
suggest you go ahead and add the capacitors to your schematic and layout.
Then as you start to bring your board on-line, experiment with the presence
or absence of the capacitors.  This is of course assuming you have that kind
of time... :-)  Student projects (any project for that matter) have a nasty
habit of consuming all available time just to get things limping along.

You mentioned that you may not understand the reason for digital and analog
grounds and why they should be connected at a single point.  Without going
into great and painful detail, ground isn't just a magic place for the
current to flow back to the power source while sitting peacefully at "0V".
It is the other half of your power distribution network and experiences all
the same fluctuations and current surges your positive (and negative)
supplies do.  Oversimplified but there it is.  If you have the time, browse
through some of the application notes found at the Linear Technology web
site also their design notes.  Especially those dealing with high bit-count
ADCs and just about anything written by Jim Williams.  Also lots of good
noise control stuff in the old Burr-Brown app notes but I'm not sure if TI
has them all on the web.  National Semiconductor and Bob Pease have lots of
good stuff too.  In short, go read.  Go read a lot!  Then build stuff, blow
stuff up and make things work.  Then try and understand why the did or
didn't work.  Become the local lab rat and take advantage of all the "free"
test equipment you have at your disposal until you graduate.

Rob Young

2005\02\24@010233 by William Chops Westfield

face picon face

On Feb 23, 2005, at 7:14 PM, John J. McDonough wrote:

> I didn't see anyplace where the price was the same, there was never
> much of a difference, and in some sizes, 1% was cheaper.  Kind of a
> surprise.

I think the manufacturing processes for surface mount resistors make it
easier to come out with 1% resistors economically, or something.

Do people in general call out tolerance on schematics for things like
pull-up resistors?  I'd normally only specify a tolerance if that
tolerance were actually required.  But then, MY collection of SMT
resistors comes from those eBay "partial reel" assortments, and my
"about 100k" resistors are actually 97.6k 1% things (and etc.)

BillW

2005\02\24@034640 by Jose Da Silva

flavicon
face
On Wednesday 23 February 2005 07:56 pm, Russell McMahon wrote:
> >> Are you actually running 2 separate gnds on your circuit board?
> >
> > Yes, analog and digital.
> >
> > Well, I _thought_ I did. :-)  What am I doing wrong? I was making
> > the assumption that digital components grounds (VSS) connect with
> > the solid DGND symbol, and analog grounds (AVSS, ground of MAX232's
> > ground, etc) connect to the standard AGND symbol. Did I make a
> > mistake or do I have a fundamental misunderstanding?

Looking at the 5v supply U4 you have it on a different ground from the PIC
chip but you have a common +5V to the chip.

Your MAX232 is on the same +5v but you should put it on the same ground as
the PIC chip. You may consider it analog, but it is a digital charge pump so
it's going to generate analog noise.

Same with U3, you are supplying +5V to digital circuits, so keep those on the
same ground as your PIC and Flash.

The ground under U1E should be the same ground as your digital PIC stuff.
These are in general all noise producing digital stuff.

U9 you have 2 grounds, yet you have the +5V wired together.
Leave your digital stuff on your digital ground, feed your analog side of the
+5V circuit through a low pass circuit. Use a resistor or inductor versus
the hardwired +5v connections between your C27/C28 and your C29.
If you really are trying to isolate digital noise, then you have to isolate
your VCCs to some degree.

U7 put all your digital stuff on 1 ground, and your analog stuff on the other
ground. I didn't look too hard, but I'm guessing that the 3.3v regulator
feeds all the analog side of the chip, therefore that would put a very good
isolation on your VCCside. If not, then you need to isolate your 3.3v
digital VCC from your analog 3.3v VCC, therefore, in this case, I would put
the regulator on the digital ground instead.

In general, you want all the noisy stuff to be spiking on the digital side
and your analog circuits on rather isolated loops. Your regulators (due to
digital circuits) and MAX232 (due to digital signal & charge pump) are going
to be rather noisy.

> As long as they connect to each other somewhere - preferably at a
> single point. If you can manage an infinite extent, zero impedance
> ground plane then you can connect to it anywhere. As you usually
> can't,  then star grounding and ground separation has its place.
> people will now howl with horror and a holy war will ensue  :-)

When talking of Analog circuits, those are grouped together with black magic.
Holy sacrifices are performed on alternate weekends.
An Altar will be provided, but please bring your own goat's blood to the
sacrifice as we usually run out of blood early.

2005\02\24@080909 by olin_piclist

face picon face
Robert Rolf wrote:
> It's a PDF. It SCALES beautifully in Adobe. 400x is quite readable,
> but you have to pan and scan to see it all.

Yeah, I could have done that.  I don't remember how you scale in Acrobat
Reader.  It's probably really easy, and with a few seconds I probably would
have found it, but at least the perceived hassle factor tipped it over to
the "---- this" side since I was doing it for free.

Neatness and care count, especially when you're asking others for a favor.
When I didn't see something that was immediately readable, my attitude was
"Why should I care if he didn't?".


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\02\24@085124 by Wouter van Ooijen

face picon face
> When I didn't see something that was immediately readable, my
> attitude was "Why should I care if he didn't?".

Zooming in to a pdf document is two mouse clicks (select the big + in
the toolbar, click on the image). If that's too much work for you, why
take the time to type a moderately lengthy reply and bother everyone
else on the list with it?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\02\24@122743 by Kenneth Lumia

picon face
>I think the manufacturing processes for surface mount resistors make it
>easier to come out with 1% resistors economically, or something.

Due to the small price differences in 1% and 5% parts, it is often the
case that it is easier to just stock the 1% parts.  Manufacturing likes
it because that's one less reel to retrieve, put into the placement
machine, remove and restock.  So the extra you pay for the part is
saved in material and labor burdens.  Purchasing is happy because
there are that fewer parts to buy.  

> Do people in general call out tolerance on schematics for things like
> pull-up resistors?  I'd normally only specify a tolerance if that
> tolerance were actually required.  But then, MY collection of SMT
> resistors comes from those eBay "partial reel" assortments, and my
> "about 100k" resistors are actually 97.6k 1% things (and etc.)

As with everything, there is a balance between too little and too
much information.  Theoretically, the BOM should always include
the specifics.  Sometimes it is easier for a tech repairing a board
to just read it off the schematic without having to take the time to
refer to the BOM.  In that case, you should call out tolerances on a
schematic.  If desired, to avoid clutter, you could create a text block
(typically on the first page) that says something to the effect: "All
resistors 5% tolerance unless otherwise specified".  Similar text
could be used with capacitors (specify dielectric).

Ken
TakeThisOuTklumiaEraseMEspamspam_OUTadelphia.net


2005\02\25@030158 by ThePicMan

flavicon
face
At 12.27 2005.02.24 -0500, you wrote:
>>I think the manufacturing processes for surface mount resistors make it easier to come out with 1% resistors economically, or something.
>
>Due to the small price differences in 1% and 5% parts, it is often the case that it is easier to just stock the 1% parts.  Manufacturing likes it because that's one less reel to retrieve, put into the placement machine, remove and restock.  So the extra you pay for the part is saved in material and labor burdens.  Purchasing is happy because there are that fewer parts to buy.

Why all the resistor I buy (rated at 5% or more) then (once measured with my ohmeter) prove to be even better than 1%, in reality?


2005\02\25@033707 by Wouter van Ooijen

face picon face
> Why all the resistor I buy (rated at 5% or more) then (once
> measured with my ohmeter) prove to be even better than 1%, in reality?

Did you measure over the full temperature range, and after (accelerated)
ageing?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\02\25@034148 by William Chops Westfield

face picon face
On Feb 24, 2005, at 9:27 AM, Kenneth Lumia wrote:

> Theoretically, the BOM should always include the specifics.

I guess that "theoretically", the BOM includes exact manufacturers and
their specific part numbers, and probably a list of approved vendors
and possible substitutes.  (sometimes, at the expense of more useful
(to tech or engineer) info like generic part number, component type, or
value.  Sigh.  Lovely to know that that resistor that burnt up was a
part number
47-123479-12b.  Thank you very much.)

What I'm wondering is if that sort of thing happens before it becomes
the responsibility of someone besides the design engineer.  A suitably
large company  has a whole "manufacturing engineering" team that
worries about such things.  I sorta assumed that before that level, I'd
be working with a distributer that would "kit" the parts for me (and do
a lot of the work of translating schematic symbols to actual 'real'
parts), and shipping them off to a contract manufacturer with "R1" type
designations or something.  Isn't that the way it works?  What level of
business wants/requires schematics laid out with exact part
designations?

BillW

2005\02\25@040637 by Robert Rolf

picon face

William Chops Westfield wrote:

> On Feb 24, 2005, at 9:27 AM, Kenneth Lumia wrote:
>
>> Theoretically, the BOM should always include the specifics.

Absolutely. Or get burned by generic parts that don't 'quite'
work. H11G2 opto for instance. Some brands are fast enough
to isolate serial comms. Others suck. You don't know until
the WRONG part is stuffed and the board fails in the field
when it warms up.

{Quote hidden}

One where the purchasing department does the buying and not the
engineer. Most clerks are clueless about part numbers, particularly
when the distributor asks 'what temperature range', 'what speed'?
I have found that nailing down the exact part number for a
production run is the hardest part of a 'quick' project.

R

2005\02\25@042932 by ThePicMan

flavicon
face
At 09.37 2005.02.25 +0100, you wrote:
>> Why all the resistor I buy (rated at 5% or more) then (once
>> measured with my ohmeter) prove to be even better than 1%, in reality?
>
>Did you measure over the full temperature range, and after (accelerated)
>ageing?

Of course not. :D


>Wouter van Ooijen

2005\02\25@073305 by Wouter van Ooijen

face picon face
> >> Why all the resistor I buy (rated at 5% or more) then (once
> >> measured with my ohmeter) prove to be even better than 1%,
> in reality?
> >
> >Did you measure over the full temperature range, and after
> (accelerated)
> >ageing?
>
> Of course not. :D

Then why are you surprised?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\02\25@084050 by Russell McMahon

face
flavicon
face
>> >> Why all the resistor I buy (rated at 5% or more) then (once
>> >> measured with my ohmeter) prove to be even better than 1%,
>> in reality?
>> >
>> >Did you measure over the full temperature range, and after
>> (accelerated)
>> >ageing?
>>
>> Of course not. :D
>
> Then why are you surprised?

Metal film "5%" resistors from a reputable manufacturer will be well
within spec across full everything range. They just can't reasonably
dumb their process down enough to get out to 5%.

This makes the process of 'select on test" whereby you found a
resistor that suited your high accuracy needs by sorting through a bin
if eg 5% values. Such a method is questionable for various reasons but
nowadays it just doesn't;t work as the R's all cluster quite tightly
about their nominal values.


       RM

2005\02\25@090340 by Gerhard Fiedler

picon face
Robert Rolf wrote:

{Quote hidden}

I find that has little to do with whether the purchasing is done by
engineers or "clerks". Nailing down the exact part number is the hardest
part because it is not trivial. If it is so difficult, no engineer can do
that without intimate knowledge of the circuit and the application and
while on the phone with a distributor... Would you know from a quick glance
at a part list or even a schematic "what speed" is required? If in doubt
the fastest? Not cost effective, and plain wrong in some cases.

In many cases something like "100n" is enough as spec... you just assume
that 10% and 50 V is good enough. But not always...

I think a staged process would make sense. Probably many bigger companies
have that, but it makes sense for smaller ones, too. And there are so many
smaller ones that setting something like this up for everybody to use would
make sense.

You have (internal) part numbers, that could look like "100n" or "100n/50V"
or whatever. These are keys into a database, where you find a list of part
with manufacturer information that fit the bill for that part. Each of
these manufacturer parts then points to a list of parts with distributor
information.

Say a generic "100n" part would point to hundreds of manufacturer parts,
all compatible with that generic part. A part like "100n/50V+" would only
point to the subset of those with 50V or more. Or a more specific part
number like "100n/50V+/Polyester" would... you get the idea.

You go in with such a part number, and the system comes back with a list of
compatible parts with complete manufacturer and distributor information. I
guess some online systems (not free) already provide the part that goes
from manufacturer to distributor parts. That would have to be coupled with
a system that goes from those internal part numbers to manufacturer part
numbers. While much of this is highly individual and depends on
preferences, I think something like an open-source effort to come up with a
system that's reasonably generic would probably be worth the effort.
Everybody could then filter the results to his/her liking ("I don't like
Maxim parts." "I never again buy at <your favorite distributor>.")

Don't know whether this is workable, but I like the idea... :)

Gerhard

2005\02\25@103528 by ThePicMan

flavicon
face
At 13.33 2005.02.25 +0100, you wrote:
>> >> Why all the resistor I buy (rated at 5% or more) then (once
>> >> measured with my ohmeter) prove to be even better than 1%,
>> in reality?
>> >
>> >Did you measure over the full temperature range, and after
>> (accelerated)
>> >ageing?
>>
>> Of course not. :D
>
>Then why are you surprised?

Because I thought that the error would be +/- 5% even at ambient
temperature, etc..


>Wouter van Ooijen

2005\02\25@114856 by Mike Hord

picon face
> >> >> Why all the resistor I buy (rated at 5% or more) then (once
> >> >> measured with my ohmeter) prove to be even better than 1%,
> >> in reality?
> >> >
> >> >Did you measure over the full temperature range, and after
> >> (accelerated)
> >> >ageing?
> >>
> >> Of course not. :D
> >
> >Then why are you surprised?
>
> Because I thought that the error would be +/- 5% even at ambient
> temperature, etc..

Also, have you tried comparing the value of resistors (same part number)
you ordered two years ago against the ones you ordered last week?

I believe homogeneity is common within a batch, but across multiple
batches, only the 5% is guaranteed.

Mike H.

2005\02\25@125828 by Kenneth Lumia

picon face

From: "William Chops Westfield" <RemoveMEwestfwspamTakeThisOuTmac.com>
>
> What I'm wondering is if that sort of thing happens before it becomes the
> responsibility of someone besides the design engineer.  A suitably large
> company  has a whole "manufacturing engineering" team that worries about
> such things.  I sorta assumed that before that level, I'd be working with
> a distributer that would "kit" the parts for me (and do a lot of the work
> of translating schematic symbols to actual 'real' parts), and shipping
> them off to a contract manufacturer with "R1" type designations or
> something.  Isn't that the way it works?  What level of business
> wants/requires schematics laid out with exact part designations?
>
--- short reply ---
It's not really a question of the amount of business.  Most schematic
software allows you to add text (such as internal part numbers) that
can be hidden on the actual schematic, but is picked up by other
software to automate tasks further down the chain.

---long reply----:)
It must be the design engineer's responsibility to generate the parts
list as only the designer knows exactly what is needed.  This is
generally done indirectly and automated as far as possible.  For
example, at a previous employer, all of the systems were linked
together.  As the designer, I would create a schematic on a Sun
Workstation using Mentor graphics schematic capture.  The
schematic software was linked to a database that held all the
parts that the company used, so in effect, I couldn't add a new part
with unknown vendors to a schematic.  If a new part was needed,
I would generate a new parts request that got every one involved,
from purchasing, production, safety, layout, etc. to verify that the
new part was acceptable.  Once approved, it could be added.
When the schematic was complete, it would be sent to the layout
people.  They would layout the board (after several go-arounds with
me for placement and general routing instructions).  Once complete,
the software would magically create a BOM with reference designators,
parts quantities, etc.  This would get sent to purchasing to
"load requirements" into the purchasing system based on projected
demand and production ramp-up.  Unless the company is really small,
purchasing is automated with the vendos/distributors.  This used to be
called EDA purchasing, now I think its under the heading B2B.  The
net effect is that the parts are automatically ordered based on lead
times for each part and humans really only need to be involved if
there is a problem (such as parts shortages). This used to be called
JIT (or JTL if you liked to complain).  As far as production goes, the
layout people send the files to production.  Production runs another
software application that generates XY coordinates and rotations
for each part.  They don't use the reference designators at all. Most
boards are too tightly packed to fit them on the silkscreen. Similar
files are sent to test engineering to build test heads for electrical test.
Even the production techs fixing the ocassional error don't normally use
the ref des.  They run on xy coordinates that feed back into a database.

Automating and using good processes goes a long way to avoiding
errors.  Then, when an error does occur, it happens on all units of that
build.  Much easier to correct than an occasional "hand stuffed" board
error.

In effect, all the above processes are driven from the designers schematic.

Ken


2005\02\25@150856 by Dave VanHorn

flavicon
face

>  so in effect, I couldn't add a new part
>with unknown vendors to a schematic.  If a new part was needed,
>I would generate a new parts request that got every one involved,
>from purchasing, production, safety, layout, etc. to verify that the
>new part was acceptable.


This is one thing that prevents me from using such packages.

Sometimes, I just want a schematic, not a whole product design with
regulatory approvals etc.
For example, I'm working on a project now, that is a "proof of concept".
Speed is far more important than anything else now, so I just slammed it out.
The only things I really needed to worry about were a few parts in unusual
packages.
For the rest, I just concentrated on getting the shape of the circuit right
(I need a resistor here) and then at the last, picked some common packages
that I knew I had in stock.  That let me go from "can we do this" to a
routed board in only a few hours.

Later, when I'm actually building it (tomorrow in this case) I'll worry
about the exact resistor and cap values, and if I need something that isn't
available in the package I picked, then I'll just fake it.

For production, I'll take a far more stringent approach, but in this case,
it's a lot like a stick, vs automatic transmission.
I can always out-drive the automatic. No matter what they do in the box, it
still doesn't know anything about the road ahead, or how I want to drive
that.  But, when I'm just getting groceries, the automatic is better.



2005\02\28@032637 by Chen Xiao Fan

face
flavicon
face
Very interesting reading. Creating the part list really takes a
lot of time for me. My company is a medium-sized company and the
schematics and layout package (P-CAD 2002) is not linked to the
ERP (Movex on IBM AS400). It is a headache to look into the
system to see the availability of part in the store and the
cost since you need to remmeber those cryptic command on the
Movex system. Just hope they could link this together. You
know we have 6-digit part numbers and they just start
to standize on nameing of the component. And they are going
to upgrade the ERP system to be web based. But it is much
slower than the old terminal based program. :(

When we want to get a new part, we need to request the part
samples, the quotations through the help of the purchasing
department. We need to creat the datasheet, ask the layouter
to create the EDA library. All need to be stored in the
engineering document manage system (EDM) and need at least
two levels of approval.

So it is true in effect, all the processes are driven from the
designers schematic (and the part list from the schematics).
But I just think this takes too much time and should put too
much administrative efforts to the design engineer. Can it
be handled at least partially by the "product engineering
department". Does someone else have a better approach?

Xiaofan

> {Original Message removed}

2005\02\28@040516 by Bob Axtell
face picon face
That's life workin' for a biiiig company....

They are being driven by ISO9002 standards... It won't change.

--Bob

Chen Xiao Fan wrote:

{Quote hidden}

>>{Original Message removed}

2005\02\28@193107 by Kenneth Lumia

picon face
Yeah, it does take time to do it right.  What happens if
the designer chooses parts that are not compatible with
the production process or purchasing lead times, etc.?  
All the designers efforts are wasted.  Getting the parts
approved first lets the designer know they are
available, cost effective and compatible.  Of course,
this process should be as fast and as automated as
possible.

The "product engineering department" can certainly
be involved in the process, however when tasks are
delegated there still needs to be someone accountable
to ensure that all tasks are completed in a timely
manner.  Project managers typically can't get down to
the level of components - they don't know the specifics -
think of it this way, do you think that your management
could juggle all your requirements as well as do their
other work?  All these other tasks would just be
placed onto other engineer's to-do list, with no specific
requirement that they drop what their doing to handle
your issues.  Of course, you don't hear that the
person hasn't even started working on the issue until
next weeks status meeting (if they attended).  
Sometimes it's just easier to do it yourself
rather than wait and hope someone
else does it in a timely and correct manner.  Don't
forget, if the design is wrong or not buildable, who
is going to get blamed?  The designer of course.

Ken
klumiaEraseMEspam.....adelphia.net

{Original Message removed}

2005\02\28@194311 by Bob Axtell

face picon face
Long, long ago, and far, far away, yours truly worked for a big company,
and designed stuff
for it.

The way it worked was there we were each given a copy of the "BIG BOOK",
which held all
the parts that the big copmpany was already buying in bulk. You could
design away to your
hearts content with those parts... until you decided you had to use
ANOTHER part. All of a
sudden the wheels fell off. It required the approval of the engineering
VP to purchase a new
part, and took 2 weeks minimum.

But it worked well, really. The company made money, and I made money,
and it is still doing
everything the same way, except that the BIG BOOK is now a proprietary
database and you
are now issued a password to get to it.

--Bob

Kenneth Lumia wrote:

{Quote hidden}

> {Original Message removed}

2005\02\28@200958 by Dave VanHorn

flavicon
face

>
>But it worked well, really. The company made money, and I made money, and
>it is still doing
>everything the same way, except that the BIG BOOK is now a proprietary
>database and you
>are now issued a password to get to it.

But sometimes, a resistor is just a resistor.  :)


2005\02\28@202001 by Kenneth Lumia

picon face

>
> But sometimes, a resistor is just a resistor.  :)
>

The keyword here is "sometimes" :)

Ken


2005\02\28@211629 by Carey Fisher - NCS

face picon face

  > >
  > > But sometimes, a resistor is just a resistor.  :)
  > >
  >
  > The keyword here is "sometimes" :)
  >
  > Ken

And then there are the "better than"s that really aren't. ;<
Carey

2005\02\28@213923 by Russell McMahon

face
flavicon
face
>> But sometimes, a resistor is just a resistor.  :)

> The keyword here is "sometimes" :)

As long as it's branded "Philips" it's just a resistor. It doesn't
matter where they make them or even who they have make them for them,
as long as they maintain their quality control.

BUT if it's double flying horse brand then a resistor can be an open
circuit, a drifting value, an unexpected breakdown at below expected
spec voltage (if it has a specd voltage), a service call to Outer
Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo / Mandalay /
Gundagai / Jindalong / Merv / Manapouri  * ... (I should be so lucky
in all cases) or a lost customer / dead car / failed deadline / ...

Poor resistors tend to "just fail" more often than good ones without
having been taken out of power / voltage / ... spec. One way to spot a
good resistor is to see if it was made by Philips. There are, of
course,  many other ways, but each has to know them for themselves.

I have no affiliation/association/financial arrangement etc etc with
Philips apart from the bond of being a happy customer.


       RM

* Can you identify where all these places are :-) ?
The river running out of Merv (aka Mary) is one of the most stunning
sights I have ever seen.



2005\02\28@222054 by Chen Xiao Fan

face
flavicon
face
Normally a company will have a list of approve vendors. This tends
to solve the issue. A resistor is not only a resistor if it
is critical component (safety related, function related).

BTW, I think most of the resistors are produced in China. So where
do you get those resistors from nowhere? :)

Xiaofan

> {Original Message removed}

2005\02\28@224223 by William Chops Westfield

face picon face
>> except that the BIG BOOK is now a proprietary database and
>>  you are now issued a password to get to it.

> But sometimes, a resistor is just a resistor.  :)

Perhaps. I wonder how many different varieties of "10k ohm resistor"
are in the book, and how many are actually needed.  Lead times on
projects are frequently long enough to get lots of brand new parts
approved, and there doesn't seem to be a lot of resistance to adding
somewhat random passive components as well.  I'll bet our "big book"
has 10k resistors in 0603, 0805, 01206, and several different multi
resistor arrays.  Sigh.  I was a bit shocked when I looked at one of
our 32 port async cards, for instance.  It had 32 instances of a
max232 type chip, each with its own 4 or so voltage multiplying caps.
Perhaps there was a good reason.  Perhaps someone just LIKED 5V only
rs232 drivers, and it was already in the "big book" as the result of
some more appropriate product (Almost everything has a console port,
for instance.)

(Of course, the reverse situation also applies.  Being unable to
upgrade the next design from 16F84 to 16F628 just because the F84
is 'already in the book' is an example that comes to mind.)

BillW

2005\02\28@233447 by Russell McMahon

face
flavicon
face
> Normally a company will have a list of approve vendors. This tends
> to solve the issue. A resistor is not only a resistor if it
> is critical component (safety related, function related).

> BTW, I think most of the resistors are produced in China. So where
> do you get those resistors from nowhere? :)

I'll just quote what I said :-)

>> As long as it's branded "Philips" it's just a resistor. It doesn't
>> matter where they make them or even who they have make them for
>> them,
>> as long as they maintain their quality control.

When it comes to resistors, Philips do it better than some. I don't
care whether it's because they are made in house or because they are
made to their specs or because they hand select and test each one
(which, of course, they don't) as long as they keep working as well as
they do.


       RM

2005\02\28@233448 by Russell McMahon

face
flavicon
face
Copied to PICList

OTHERS - read comment on ceramic caps towards end.


> Sorry to burst your bubble but Philips do not make resistors and have not
> done so for some time.

No bubble.
As I said:
 > As long as it's branded "Philips" it's just a resistor. It doesn't
 > matter where they make them or even who they have make them for them,
 > as long as they maintain their quality control.
> The resistive products division of Philips (along with many other of their
> business units) were sold off to various parties (in many cases via
> managment buyouts) in recent years.
>
> The products continue to be made however (resistors by Phicomp  - A Yageo
> company), and have the same quality standards and even the same 12NC part
> numbers, and in many cases are available from the same distribution
> channels
> as the original Philips components.

Meets my criteria :-)


{Quote hidden}

2005\02\28@234633 by Dave VanHorn

flavicon
face
At 10:42 PM 2/28/2005, William "Chops" Westfield wrote:
>>>except that the BIG BOOK is now a proprietary database and
>>>  you are now issued a password to get to it.
>
>>But sometimes, a resistor is just a resistor.  :)
>
>Perhaps. I wonder how many different varieties of "10k ohm resistor"
>are in the book, and how many are actually needed.

Well, if I'm doing a schematic for production, that's one thing, and I want
exact parts unless I say otherwise.
If I'm doing a prototype, or just a schematic to illustrate or document
something, then a resistor is just a resistor.

As I said, most of my designs start out as "just a schematic" to get the
shape of the circuit right.
I don't want to worry about 1%, 5% yaego or Panasonic, I just want a
resistor there. I may not even know the value closer than an order of
magnitude, so it just slows me down to have to go specify an exact part.

Sort of the schematic version of "Measure with a micrometer, Cut with a
hacksaw".





'[EE] Pull-up/down resistors'
2005\03\01@013110 by Bob Barr
flavicon
face
On Mon, 28 Feb 2005 23:46:51 -0500, Dave VanHorn wrote:


>
>Sort of the schematic version of "Measure with a micrometer, Cut with a
>hacksaw".

That sounds like a variation of the old joke about mil-spec:

Measure with micrometer; mark with chalk; cut with ax; spend three
days writing documentation. (Or maybe that was ISO 9000.)


Regards, Bob

2005\03\01@045009 by steve

flavicon
face
> As long as it's branded "Philips" it's just a resistor.

Not a good thing if you've specified NFR25 or FRC01 series.
They are supposed to be a fuse too.

> > The result is often an overheating regulator (or one
> > which thermally cycles over 10's of seconds)

Ah. The thermal oscillator. I hate it when that happens. :-)

Steve.

==========================================
Steve Baldwin                          Electronic Product Design
TLA Microsystems Ltd             Microcontroller Specialists
PO Box 15-680, New Lynn                http://www.tla.co.nz
Auckland, New Zealand                     ph  +64 9 820-2221
email: RemoveMEsteveEraseMEspamEraseMEtla.co.nz                      fax +64 9 820-1929
=========================================



2005\03\01@071838 by Russell McMahon

face
flavicon
face
>> As long as it's branded "Philips" it's just a resistor.

> Not a good thing if you've specified NFR25 or FRC01 series.
> They are supposed to be a fuse too.

The point was MEANT to be that Philips resistors tend to meet specs
and have specs to meet. Even if they are no longer made by Philips :-)


       RM

2005\03\01@075232 by olin_piclist

face picon face
Russell McMahon wrote:
> Outer
> Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo / Mandalay /
> Gundagai / Jindalong / Merv / Manapouri ...
>
> Can you identify where all these places are :-) ?

I'm not sure I know where the last 5 are off the top of my head, but I get
the impression you were trying to pick out of the way or hard to reach
places.  If so, Kalamazoo doesn't belong.  It's in southern Michigan on
interstate 94, about half way between Detroit and Chicago.  I just doesn't
seem up there with Outer Mongolia or any place in Uzbekistan.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\03\01@100912 by Rob Young

picon face

> Russell McMahon wrote:
>> Outer
>> Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo / Mandalay /
>> Gundagai / Jindalong / Merv / Manapouri ...
>>
>> Can you identify where all these places are :-) ?
>
> I'm not sure I know where the last 5 are off the top of my head, but I get
> the impression you were trying to pick out of the way or hard to reach
> places.  If so, Kalamazoo doesn't belong.  It's in southern Michigan on
> interstate 94, about half way between Detroit and Chicago.  I just doesn't
> seem up there with Outer Mongolia or any place in Uzbekistan.
>
>

He could also substitute Yazoo (put a resistor manufacturing plant at the
head of the river and their slogan could be "We have resistors up the
Yazoo") for Kalamazoo.  Or perhaps Tippecanoe (but not Tyler Too).

Rob

2005\03\01@102717 by Joe McCauley

picon face
Well it would be almost as hard for me to reach Kalamazoo as
Mandalay......:)


Russell McMahon wrote:
> Outer
> Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo / Mandalay /
> Gundagai / Jindalong / Merv / Manapouri ...
>
> Can you identify where all these places are :-) ?

I'm not sure I know where the last 5 are off the top of my head, but I get
the impression you were trying to pick out of the way or hard to reach
places.  If so, Kalamazoo doesn't belong.  It's in southern Michigan on
interstate 94, about half way between Detroit and Chicago.  I just doesn't
seem up there with Outer Mongolia or any place in Uzbekistan.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\03\01@105240 by Spehro Pefhany

picon face
At 07:52 AM 3/1/2005 -0500, you wrote:
>Russell McMahon wrote:
>>Outer
>>Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo / Mandalay /
>>Gundagai / Jindalong / Merv / Manapouri ...
>>
>>Can you identify where all these places are :-) ?
>
>I'm not sure I know where the last 5 are off the top of my head, but I get
>the impression you were trying to pick out of the way or hard to reach
>places.  If so, Kalamazoo doesn't belong.  It's in southern Michigan on
>interstate 94, about half way between Detroit and Chicago.  I just doesn't
>seem up there with Outer Mongolia or any place in Uzbekistan.

Depends on your perspective I guess. Mandalay is an exotic Asian cultural
center  with bustling markets, a faded colonial past, and almost a million
inhabitants, Kalamazoo is a somewhat decrepit and obscure industrial town
in the American midwest (population ~70K, former home of the Checker car
company), a possible rest stop on a car trip from Detroit or Toronto to
Chicago on I-94.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
RemoveMEspeffspam_OUTspamKILLspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com




2005\03\01@144037 by Peter L. Peres

picon face


On Tue, 1 Mar 2005, Russell McMahon wrote:

> As long as it's branded "Philips" it's just a resistor. It doesn't matter
> where they make them or even who they have make them for them, as long as
> they maintain their quality control.
>
> BUT if it's double flying horse brand then a resistor can be an open circuit,
> a drifting value, an unexpected breakdown at below expected spec voltage (if
...

Upon lesipt of you honolab msg wee now sip "Pilipps" bland palts for yor
pleasar

2005\03\01@183413 by Russell McMahon

face
flavicon
face
>> Outer
>> Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo /
>> Mandalay /
>> Gundagai / Jindalong / Merv / Manapouri ...
>>
>> Can you identify where all these places are :-) ?
>
> I'm not sure I know where the last 5 are off the top of my head, but
> I get
> the impression you were trying to pick out of the way or hard to
> reach
> places.  If so, Kalamazoo doesn't belong.  It's in southern Michigan
> on
> interstate 94, about half way between Detroit and Chicago.  I just
> doesn't
> seem up there with Outer Mongolia or any place in Uzbekistan.

Depends where you live :-)

Manapouri is in my country and accessible enough to me (needs an
interisland crossing or an air hop).
But kalamazoo is to me the name of a business system. I've looked it
up on maps but never been there. I have, however, flown over Merv.

Mandalay - "... On the road to Mandalay, where the flying fishes play,
and the sun comes up like thunder out of China 'cross the bay."
Gundagai  - " .. coz the dog sits on the tuckerbox, 6 miles from
Gundagai." - and it still does - set in ?bronze for all time. I've
seen it.
Jindalong - Dancing in the desert - Australian outback over the
horizon radar installation. Never been there.
Merv - or "Mary" / Murgab. Oasis with river flowing north to a sandy
death in the Karkhoram desert. On the old silk route. Viewed by me and
no other passengers at 4 am from 35,000 feet courtesy of Mr Boeing.
Manapouri - see above.

Kalamazoo. Somewhere in the trackless wastes of Southern Michigan.
Never been there.


       RM




2005\03\01@190634 by Dave VanHorn

flavicon
face

>
>Kalamazoo. Somewhere in the trackless wastes of Southern Michigan. Never
>been there.

Indeed, it's a couple hours north of me.  I'm outstanding in my cornfield.


2005\03\02@095630 by Howard Winter

face
flavicon
picon face
Bob,

On Mon, 28 Feb 2005 22:31:21 -0800, Bob Barr wrote:

> That sounds like a variation of the old joke about mil-spec:
>
> Measure with micrometer; mark with chalk; cut with ax; spend three
> days writing documentation. (Or maybe that was ISO 9000.)

LOL!  A former member of a team I ran had trained as a Civil Engineer - he told me how they design bridges.  
Take the worst imaginable circumstances (100-year wind at exactly the wrong angle, the downwind roadway loaded
with heavy high-sided vehicles, the upwind roadway empty, snow lying on the decks, and so on) and calculate
all the stresses to 4 decimal places.  Work out how they all add up (to 4 decimal places) and calculate the
total strength needed to withstand them without damage.  Then treble it!  :-)

Cheers,


Howard Winter
St.Albans, England


2005\03\02@101323 by Howard Winter

face
flavicon
picon face
Russell,

On Wed, 02 Mar 2005 12:33:42 +1300, Russell McMahon wrote:

> Outer Mongolia / Uzbekistan / Pago Pago / Timbuktoo / Kalamazoo /
> Mandalay / Gundagai / Jindalong / Merv / Manapouri ...
>
> Can you identify where all these places are :-) ?

Let's see:

Outer Mongolia:  North-East of the Eurasian continent (to the right of Siberia, perhaps?)
Uzbekistan: South-East of the former USSR, near (North of?) Afghanistan
Pago Pago:  I heard you the first time!  :-)  Guessing: an island in Indonesia?
Timbuktoo:  Africa, about half-way down the left side, I think
Kalamazoo: Somewhere in the Midwest of the USA
Mandalay: Far East, on the coast, near China
Gundagai, Jindalong, Merv: Never heard of these
Manapouri: Been there!  South Island New Zealand, near the South-West corner, a lake with a hydro-electric
plant that's deep underground, supplying electricity to an aluminium smelting plant, with any excess being
used by the local towns.  (I hope that's the place - it's been a couple of years).  Amazing bus ride down a
2km curving tunnel to the plant, I don't think I've ever been so deep inside the Earth!

Cheers,



Howard Winter
St.Albans, England


2005\03\02@112101 by olin_piclist

face picon face
Howard Winter wrote:
> Outer Mongolia:  North-East of the Eurasian continent

That would be rather wet.  Outer Mongolia is still IN Asia.

> (to the right of Siberia, perhaps?)

Again wet, since Siberia extends all the way to the Pacific ocean.

> Uzbekistan: South-East of the former USSR,

Actually IN the former USSR.  It was one of the 15 republics of the Soviet
Union.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\03\05@221850 by Mike Singer

picon face
Olin Lathrop wrote::
> Howard Winter wrote:
> > Outer Mongolia:  North-East of the Eurasian continent

The most North-East of the Eurasian continent is Chukotka.
The Governor of Chukotka bought Chelsea not so long ago.

From
sports.yahoo.com/sow/ news?slug=afp-fblenglcupchelsea&prov=afp&type=lgns

"A week of turmoil ended in triumph for Chelsea as they came from
behind to beat Liverpool and win the League Cup"

> That would be rather wet.  Outer Mongolia is still IN Asia.
>
> > (to the right of Siberia, perhaps?)
>
> Again wet, since Siberia extends all the way to the Pacific ocean.

If I'm not mistaken Siberia doesn't. The ajacent to Pacific ocean part
of Russia consists of Chukotka, Kamchatka, Yakutiya, Amurski krai.
Yakutiya contains the so called Kolyma the most fearful part of
Archipelag Gulag.

> > Uzbekistan: South-East of the former USSR,
>
> Actually IN the former USSR.  It was one of the 15 republics of the Soviet
> Union.

Uzbekistan: Central-South of (not off:-) the former USSR, one of its
15 republics.

Mike.

2005\03\06@151052 by olin_piclist

face picon face
Comrade Singer wrote:
> If I'm not mistaken Siberia doesn't. The ajacent to Pacific ocean part
> of Russia consists of Chukotka, Kamchatka, Yakutiya, Amurski krai.
> Yakutiya contains the so called Kolyma the most fearful part of
> Archipelag Gulag.

OK, you should know this better than me.  I thought Kamchatka, the Kuril
Islands, and region around Vladivostok were part of Siberia, but I was
apparently wrong about that.  We only have old maps around here that just
have "Evil Empire" stamped in bold letters accross central Asia ;-)

>> Actually IN the former USSR.  It was one of the 15 republics of the
>> Soviet Union.
>
> Uzbekistan: Central-South of (not off:-) the former USSR, one of its
> 15 republics.

Um, I thought that's what I said.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\03\06@201538 by Mike Singer

picon face
part 1 1217 bytes content-type:text/plain; charset=US-ASCII (decoded 7bit)

Olin Lathrop wrote:
> Comrade Singer wrote:
> > If I'm not mistaken Siberia doesn't. The ajacent to Pacific ocean part
> > of Russia consists of Chukotka, Kamchatka, Yakutiya, Amurski krai.
> > Yakutiya contains the so called Kolyma the most fearful part of
> > Archipelag Gulag.
>
> OK, you should know this better than me.  I thought Kamchatka, the Kuril
> Islands, and region around Vladivostok were part of Siberia, but I was
> apparently wrong about that.  We only have old maps around here that just
> have "Evil Empire" stamped in bold letters accross central Asia ;-)

I see, should you have Internet access, you  would probably check
Wikipedia first
Posting by snail-mail is so boring. Oxcarts are moving slooow ;-)

And my mistake - Yakutiya is still a part of Siberia.


> >> Actually IN the former USSR.  It was one of the 15 republics of the
> >> Soviet Union.
> >
> > Uzbekistan: Central-South of (not off:-) the former USSR, one of its
> > 15 republics.
>
> Um, I thought that's what I said.


Perhaps my language problem, since I thought that's what the OP said

Mike.


part 2 3042 bytes content-type:image/gif; name="Sergey" (decode)


part 3 35 bytes content-type:text/plain; charset="us-ascii"
(decoded 7bit)

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...